Open Source Integrated Library System

Evergreen on IRC

#evergreen Logs for Wednesday, October 13th, 2010

< Tuesday, October 12th, 2010Raw Log FileThursday, October 14th, 2010 >
#TimeNickMessage
#05:10:17JMCraig has quit IRC
#07:39:32sfortin has joined #evergreen
#07:44:44collum has joined #evergreen
#07:44:50sfortin has quit IRC
#07:45:12r123 has joined #evergreen
#07:45:18granitize has joined #evergreen
#07:45:28sfortin has joined #evergreen
#08:00:44collum has left #evergreen
#08:11:48StephenGWills has joined #evergreen
#08:25:04collum has joined #evergreen
#08:52:53phasebb has joined #evergreen
#09:02:16Dyrcona has joined #evergreen
#09:03:25phasebb has quit IRC
#09:08:34Meliss has joined #evergreen
#09:17:23bshum has joined #evergreen
#09:18:47mrpeters-isl has quit IRC
#09:32:48jenny has joined #evergreen
#09:35:51yboston has joined #evergreen
#09:42:15mrpeters-isl has joined #evergreen
#09:45:19rsinger has quit IRC
#09:52:48rsinger has joined #evergreen
#10:27:13mrpeters-isl has quit IRC
#10:27:22mrpeters-isl has joined #evergreen
#10:34:18dave-esi has quit IRC
#10:49:55dave-esi has joined #evergreen
#11:16:01kmlussier has joined #evergreen
#11:20:36afterl has joined #evergreen
#11:22:08r123 has quit IRC
#12:41:53tsbere is amazed how short power cables come, but also how annoying it is to find some shorter ones locally
#12:43:17mrpeters-isl has quit IRC
#12:43:22mrpeters-isl has joined #evergreen
#12:43:44mrpeters-isl has joined #evergreen
#12:43:45mrpeters-isl has left #evergreen
#12:43:53mrpeters-isl has joined #evergreen
#12:44:23atz__tsbere: just get a UPS w/ a long cable
#12:45:27tsbereatz__: I don't want a long cable. I have plenty of long cables. And a big honking rack-mounted UPS for that matter complete with power plugs in the rack close to where the servers rack in. I want short cables that don't have several feet of excess hanging out before being plugged into the UPS.
#12:46:21atz__oh, i thought you were complaining the local ones were too short.
#12:47:45tsbereNo, the nearest supplier sells 6 to 15 footers. I need maybe 4.
#12:48:04csharp finally gets around to installing EG-alpha4 on i386 Ubuntu Lucid
#12:48:12csharpno problems so far
#12:49:09tsbereAlthough you have to admit, they do come really really short. Found one that is so short you can't bend it.
#12:49:32bshumYay csharp
#12:50:02csharpI actually attempted an install on CentOS in a VM - no success
#12:50:06bshumcsharp: Out of curiosity, when you're doing your tests, how much RAM are you assigning to them? (I'm assuming you're doing them on VM)
#12:50:20csharpbut that's not an environment I'm that familiar with
#12:50:42csharpbshum: on my VMs, it's usually 1 GB
#12:50:45bshumcsharp: I oddly ran out of RAM the other day installing EG trunk, only assigned 1 GB to it.
#12:51:14csharpbshum: on this (metal) machine, it's 3GB
#12:51:35bshumcsharp: Just curious. I usually go with 1GB, but I thought I'd check :)
#12:51:44csharpbshum: I haven't run out before
#12:51:48csharpinteresting
#12:52:15csharpI *did* get OpenSRF 1.6 running on CentOS, though, for those interested
#12:52:36mrpeters-islRC4 eh? looks like i need to get on upgrading from RC2 :)
#12:52:39csharpthe Makefile.install for Evergreen is what failed - I'd have to do it again to get specific
#12:53:06mrpeters-islstill have been meaning to try this on Squeeze x64 too
#12:53:55csharpmrpeters-isl: Squeeze stills screws up GRUB2 for me in VMs - I've yet to have a clean installation experience with it
#12:54:19mrpeters-isloh yeah? hmm x64 or both?
#12:54:30csharpboth - but lately x64
#12:54:33mrpeters-islvery interesting
#12:54:34jeffwhat VM tech?
#12:54:39mrpeters-isli have it working on Squeeze
#12:54:41csharpjeff: VirtualBox
#12:54:42mrpeters-islbut bare metal
#12:55:10mrpeters-islwhat happens to grub2? i've gotten to be quite friendly with grub2 in working with Xen lately
#12:55:28csharpmrpeters-isl: it fails during update-grub somehow
#12:55:50csharpI've never had the patience to track it down, since I know lenny works ;-)
#12:55:51mrpeters-islcsharp: see, with xen...it didn't write the /boot/grub/grub.cfg file right
#12:56:01mrpeters-island you had to manually make some modifications
#12:56:06mrpeters-isli'd be curious to see what that file looks like
#12:56:23csharpmrpeters-isl: hmm - I've not used Xen before
#12:56:24mrpeters-isli dont know if that was just a bug with the xen kernel (that would maybe be there in other virtualization techs?)
#12:56:39csharpour admins here are big KVM people
#12:57:03mrpeters-islyeah. you have me curious now. maybe ill try my hand at installing EG in a squeeze DomU
#13:12:10lisppastecsharp pasted "PostgreSQL errors" at http://paste.lisp.org/display/115506
#13:12:45tsbereI am not sure those are an issue. <_<
#13:12:51csharp^^ got these errors when running the eg_db_config.pl script
#13:13:10tsbereIf you are concerned, though, run eg_db_config.pl again and see if they still show up
#13:13:25csharptsbere: will do - just in case - thanks
#13:13:43tsbere thinks it is more of a "we drop these before creating them for sanity purposes, and they aren't there in a clean DB" thing
#13:14:18eeeviltsbere: it is
#13:16:09csharptsbere: eeevil - thanks - no errors on the second run - now I see why
#13:23:12jamesrf has joined #evergreen
#13:30:57lisppastecsharp pasted "Apache2 log errors" at http://paste.lisp.org/display/115507
#13:32:44csharp seeks SRU/Request.pm...
#13:34:08csharpI must've missed a CPAN error methinks...
#13:36:59csharpweird... - it didn't install...
#13:37:15csharpshouldn't the Makefile script stop if a dependency is not met?
#13:37:33csharp(Makefile.install to be more precise)
#13:41:55eeevilno for cpan failures ... it's currently "outside" make's perview
#13:45:28tsberecpan also has this annoying issue wherein I believe it returns 0 to the shell regardless of whether it did anything. The only time it doesn't is if cpan itself crashed or errored out.
#13:47:58eXcAliBuR has joined #evergreen
#13:48:15eXcAliBuRi get this error when trying to log in... Error Communicating with Authentication Server
#13:48:19eXcAliBuRwhat am i missing?
#13:48:22eXcAliBuRis something not running?
#13:52:14tsbereeeevil: From yesterday, you mentioned something about it would be good if I could get auto update working with matchpoints instead of circ rules, what did you mean by that?
#13:52:27tsbereeXcAliBuR: Did you restart apache after running osrf_ctl.sh?
#13:53:21eXcAliBuRno...
#13:53:26eXcAliBuRi restarted the whole server
#13:53:55tsbereeXcAliBuR: Once you run osrf_ctl.sh to start everything (-l -a start_all or -a start_all depending on configuration) you need to restart apache.
#13:54:43eXcAliBuRok did that and same error
#13:56:14tsbereAlso, if you had run osrf_ctl.sh with a start_all and then didn't run it with a stop_all before shutting down or restarting the entire server you likely need to clear out /openils/var/run/*.pid and re-run the start_all variant.
#13:56:21tsbere(and then restart apache again)
#13:59:09eXcAliBuRthat worked
#13:59:10eXcAliBuR:}
#13:59:12eXcAliBuRthank you
#13:59:38eXcAliBuR has quit IRC
#14:01:19parsr has joined #evergreen
#14:06:41parsreeevil: saw your reference to 1.6.2 in changeset 18308. I'm planning another upgrade cycle this weekend - any chance 1.6.2 to be cut soonish?
#14:20:16jenny has quit IRC
#14:24:00mrpeters-islOpenSRF 1.6 installed without a hitch on Debian Squeeze x64 (October 11th build)
#14:24:06mrpeters-islonto Evergreen 2.0 RC4
#14:25:45mrpeters-islthere is some strange liquid falling from the sky in Indiana....interesting...
#14:30:33gdunbar has quit IRC
#14:38:13eeevilparsr: tentatively, 1.6.2 would be nearly concurrent with 2.0 gold
#14:40:52jenny has joined #evergreen
#14:40:59jenny has left #evergreen
#14:42:04parsreeevil: ok, very good. just want to double check so that I don't do the upgrade day before next release. There's the need to upgrade versus the "wait for next dot release" with more good stuff angle to assess (especially for me since I'm always pressed for blocks of time and have a number of OPAC and other customizations that I need to check against each upgrade).
#14:44:10bshumparsr: Out of curiosity, which version are you upgrading from/to?
#14:47:18mrpeters-islEG 2.0 RC4 working like a charm on x64 squeeze as well
#14:47:27parsrbshum: 1.6.0.7 to 1.6.1.2; performed it on our our dev server earlier this week, so feel more comfortable now that I overcame a bunch of unrelated issues (my errors following an OS upgrade from 8.04 to 10.04 LTS that added a few twists to my own goof ups).
#14:48:13parsrbshum: to clarify, this will be a prod server upgrade
#14:48:17bshumparsr: Aha, nice. We're hoping to make that leap someday from 1.6.0.2
#14:48:38bshumI wish you luck then.
#14:52:17mrpeters-islanyone gotten brave enough to do an upgrade from 1.X > 2.0 RC4 yet?
#14:55:29parsrmrpeters-isl: I think Robert Soulliere did. See post on dev list re: "Evergreen 2.0 alpha4 new sidebar facets" today or yesterday
#14:56:47tsbere likes placing orders for cables and having them in a little over an hour :D
#14:58:48mrpeters-islI think i may have discovered a bug...when i try to retrieve a patron by barcode and it isnt found, i can't clear out the field and search again
#14:58:55mrpeters-isli have to reload the retrieve patron by barcode interface
#14:59:51mrpeters-islrather, i can "CTRL+A" to clear out that field, but i cant place a cursor in it to correct a mistake
#15:00:05StephenGWills has quit IRC
#15:00:06mrpeters-isl**CTRL+A and backspace
#15:00:49tsberemrpeters-isl, usually I see that it clears the field for me. I can double check in one of the installs I have sitting here, though.
#15:01:35mrpeters-islperhaps it is just me...it caught me off guard becuase i can usually place a cursor and fix the one digit ive missed
#15:02:21tsbereOk, correction, for me it literally pre-CTRL+A's me, puts the cursor in there, and I can then change the cursor position to add/remove/fix digits
#15:02:35mrpeters-islok. ignore me then. perhaps its something on my machine
#15:03:16tsberenote that neither my server side nor staff client side installs are 100% up to date with trunk or rel_2 or any tags right now, though. Something may have been introduced since I did a full build.
#15:03:41mrpeters-isldo password reset emails derive from a template, such as hold_notification_template.example?
#15:04:11mrpeters-islor is that hard coded elsewhere...just curious to see the text of one
#15:19:39eeevilmrpeters-isl: it's an A/T template, IIRC
#15:20:17mrpeters-isleeevil: is that similar to the xsl sheets for printed notices?
#15:20:36mrpeters-isli dotn know the proper name for those?
#15:21:31eeevilmrpeters-isl: no, it's under the notifications interface ... if I'm remembering where dbs put it correctly
#15:22:18mrpeters-islah excellent i will have a look
#15:27:08pmplett has joined #evergreen
#15:40:45Meliss has quit IRC
#15:41:22mrpeters-islno dice on the A/T template for Password Reset. At least not in the RC4 release...
#15:43:50Meliss has joined #evergreen
#15:46:12eeevildenials: help us dbs-wan, you're our only hope!
#15:47:49tsbereeeevil: Still waiting on answer from question earlier. Did you miss it, or decline to answer?
#15:48:02eeevilI missed it
#15:48:13eeevilahh
#15:48:15eeevilI see it
#15:49:54eeevilsec
#15:50:21phasefxmrpeters-isl: you found the template and changing it didn't work, or you didn't find the template?
#15:50:25afterlSince eevil referenced Star Wars a couple of minutes ago, and since this proved to be a hit when I posted it on my FB page (by some of you people, I might add)...
#15:50:28afterlhttp://laughingsquid.com/the-making-of-star-wars-the-empire-strikes-back/
#15:51:00afterlOh, sorry, I meant eeevil.
#15:51:11afterlNot sure how important that third 'e' was.
#15:52:53eeeviltsbere: so, when active_date is passed, an A/T event will be generated and will update the appropriate duration rule with the new ceiling
#15:53:07eeeviltsbere: however, I was thinking more about what you suggested
#15:53:19eeevilI don't think that moving it to the matchpoint would be better
#15:53:34eeevilit would be different, but not better
#15:55:20eeevilcurrently, you'll have X duration rules per same-ceiling. but in practice that means, say 2 -- in the academic example, one for students and one for staff
#15:55:36tsbere hadn't figured out a number of details about how hard due dates were in there yet
#15:56:02eeevilactually, I take that back, it won't be an A/T event
#15:56:09eeevilit'll be a nightly jobs, IIRC
#15:56:18eeevilbut either way ...
#15:56:38mrpeters-islphasefx: didnt find the template
#15:56:47mrpeters-isllooking in the staff client action triggers
#15:58:13parsr has quit IRC
#15:58:13eeevilso, you have to say "this thing [duration rule or matchpoint] uses this /stream/ of ceilings"
#15:58:14mrpeters-isli have lots of notifications for acq, and so on...but not one for the password reset instructions
#15:59:10eeevilwith a duration rule you're already dealing with how to set the due date -- now + short/normal/long, based on the copy setting
#15:59:35eeevilso it's a natural place to put a "but not past date X" rule
#15:59:43Meliss has quit IRC
#16:00:02tsbereeeevil: Right now, I am not 100% sure (I have been looking at other things today), but it looks like something is seriously wrong with the picture. If you have different durations for different kinds of items it looks like you need different hard due date settings for each duration rule.
#16:00:05eeeviland it's easy to update, because duration rules are stable -- they don't get tweaked and adusted all the time, blown away and recreated, etc
#16:00:49eeevilthat's no different than matchpoints
#16:02:09eeeviltsbere: can you give a concrete example that would be used in practice?
#16:02:12tsbereeeevil: I would say that the better option would have been to have the relation go the other way, instead of the hard due date referring to a duration rule, have the duration rule (or matchpoint) refer to a hard due date. Then each institution has a hard due date (if you have more than one) that is updated in one place for all rules (duration or matchpoint).
#16:02:39eeeviltsbere: but each institution does not have /A/ hard due date
#16:03:08eeeviland, the hard due date is to roll over automatically
#16:04:11tsbereeeevil: Ok, lets assume that we are talking just students. Books are lent out for 3 weeks, DVDs for 1 week, and they have a set of books on a course circ modifier that go out for 2 weeks. That is 3 sets of duration rules, and I would assume all would get the same hard due date, that now needs to be set in 3 places.
#16:04:36phasefxmrpeters-isl: it's labelled 'Password reset request notification'. May have to use the obscure "Next" link to page through the templates
#16:05:28tsbereeeevil: Also, with fallthrough (which I keep looking at things from) a single matchpoint rule could be applied to all students to set a hard due date regardless of duration, a different one for faculty (if any), etc
#16:05:38eeeviltsbere: well, "in 3 places" is a bit disingenuous ;) ... you look at the duration rules that you care about and set them all up for the next 2 semesters (because that's what you have dates for) and you're done in 5 minutes
#16:06:55mrpeters-isl proceeds to hide in a corner
#16:07:14mrpeters-islthanks phasefx i wasnt expecting that there were so many in there...never even noticed the next
#16:07:17mrpeters-islim an idiot...sorry
#16:07:37phasefxmrpeters-isl: happened to me too, or I wouldn't have known to suggest it
#16:08:08mrpeters-islreally nice to see so many in here now though!
#16:08:18mrpeters-islnow i just can't get the emails to be sent out :(
#16:08:32mrpeters-islwhich, makes total sense now
#16:08:35mrpeters-islthat i see its disabled!
#16:09:43sfortin has quit IRC
#16:10:08tsbereeeevil: Of course, with fallthrough, every single setting that is tied outright to another makes fallthrough less useful, so seeing new settings being tied to old ones annoys me. <_<
#16:11:01granitize has quit IRC
#16:11:09eeeviltsbere: I'm not saying an alternate design shouldn't be persued ... but (and this is a broader suggestion -- feel free to take it with a grain of salt) I'd really like to see the basic fall-through stuff (perhaps even with renewal support, but only if it includes the backstopping renewal field on the duration rule at first) get in really quickly ... I want to see how that works out, how in practice it makes things better
#16:11:19eeeviland /then/ start foldling other stuff into it
#16:12:51eeevilIOW, I think you have about 2-3 revisions worth of ideas in one patch ATM ... that's not a slight -- you're just too fast for us to give anything a proper looking over! (with the push for 2.0, most of the committers are swamped, as you know)
#16:14:36eeevilwould you be open to breaking the patch up into, say, 1) fallthrough without renewal support 2) renewal fallthrough on top of that 3) due-date ceilings on matchpoints (note we still need automatible rollover support)
#16:14:40eeevil?
#16:14:57tsbereeeevil: #1 is still posted on launchpad, I believe.
#16:15:08atheos has left #evergreen
#16:15:54eeevilok, good, perfect ... I really want to see that get in, but I until 2.0 is at least mostly through beta I can't promise it the attention it deserves
#16:15:57tsbereI still haven't looked at the due-date ceilings beyond trying to figure out how it was implemented
#16:16:13tsbere(and not liking what I have seen already)
#16:17:33eeevilunderstood. fwiw, I generally agree with the premise of breaking up the tied bits. but I fear the upgrade path, and the changing of expectations, and the breaking of legacy code ... so, there will always be some slack time when we have to try to support the old and the new
#16:18:44tsbereeeevil: I understand that. I don't like tieing new things in to old ones at this stage, though.
#16:19:48eeevilthe ceilings are hopefully a last gasp of the circ-script mentality -- but because the code leans in that direction (in the biz-logic, I mean) it's also still the shortest path (aka, least cost) to implementation ... what you're working on with fallthrough, and pushing for the ideas behind it, is setting us up to cut that cord
#16:20:23eeevilyou're making it easier to not do that again in the future, IOW ... :)
#16:20:53tsbereahhh. That is a good thing.
#16:21:14bshum has quit IRC
#16:21:16collum has left #evergreen
#16:21:16tsbereI will have to see about figuring out the best way to handle the hard due dates sometime in the next week or two.
#16:21:29eeeviltsbere++
#16:26:56finnapz has quit IRC
#16:27:15finnapz has joined #evergreen
#16:27:48mrpeters-islanyone opposed to me "dressing up" the password reset form to look just as nice as craftsman/default opacs do?
#16:28:11mrpeters-isljust with some simple edits to reset-form.tt2
#16:28:37eeevilmrpeters-isl: I want it UGLY. ... er, no, send a patch :)
#16:28:54phasefxuse lots of blink tags
#16:29:06eeeviland a <marquee>
#16:29:20phasefxflash navigation
#16:29:30tsbereNested marquees in blink tags!
#16:29:44atz__ seizure
#16:30:07eeevilfull-page, mouseover popups!
#16:31:37phasefxanimated gifs like this one: http://celebs.icanhascheezburger.com/2010/10/12/celebrity-pictures-infinite-sadness/
#16:32:40eeevilthat's nice
#16:37:26tsbere thinks he knows how hard due date functionality is intended to work
#16:37:49tsbereNow I can start planning out all the problems I think are there and come up with a solution. <_<
#16:38:55afterl has quit IRC
#16:41:36senatorld
#16:41:56senator whistles innocently
#16:42:00berickno input files found
#16:42:04berickdoh, too late
#16:44:18tsbereSo, after minimal thinking about it, I think I came up with a workable solution already that makes most of my issues with the current hard due dates go away (well, beyond being tied to duration rules..........)
#16:46:58tsbereneeds one more table, though. :(
#16:48:00DyrconaAlways need one more table.....One more table and my database will be perfectly in 5th normal form! :p
#17:07:17jenny has joined #evergreen
#17:18:50kmlussier has quit IRC
#17:22:39jamesrf has quit IRC
#17:30:56Dyrcona has quit IRC
#17:43:13pmplett is now known as pmpafk
#17:44:11jamesrf has joined #evergreen
#18:05:10yboston has quit IRC
#18:17:50LBA has joined #evergreen
#19:13:24phasebb has joined #evergreen
#19:14:38pmpafk is now known as pmplett
#19:35:50LBA has left #evergreen
#19:51:58jamesrf has quit IRC
#20:05:56jenny has quit IRC
#20:37:09jamesrf has joined #evergreen
#20:54:30pmplett is now known as pmpafk
#20:59:14brendan2 has joined #evergreen
#22:26:51brendan2 has quit IRC
#22:53:22phasebb has quit IRC
< Tuesday, October 12th, 2010Raw Log FileThursday, October 14th, 2010 >